dynamics of the lake michigan food web1970-2000

8/13/2019 Dynamics of the Lake Michigan Food Web1970-2000

1/18

PERSPECTIVE

Dynamics of the Lake Michigan food web,

19702000

Charles P. Madenjian, Gary L. Fahnenstiel, Thomas H. Johengen,

Thomas F. Nalepa, Henry A. Vanderploeg, Guy W. Fleischer,

Philip J. Schneeberger, Darren M. Benjamin, Emily B. Smith, James R. Bence,

Edward S. Rutherford, Dennis S. Lavis, Dale M. Robertson, David J. Jude, and

Mark P. Ebener

Abstract: Herein, we document changes in the Lake Michigan food web between 1970 and 2000 and identify the fac-tors responsible for these changes. Control of sea lamprey (Petromyzon marinus) and alewife (Alosa pseudoharengus)populations in Lake Michigan, beginning in the 1950s and 1960s, had profound effects on the food web. Recoveries oflake whitefish (Coregonus clupeaformis) and burbot (Lota lota) populations, as well as the buildup of salmonine popu-lations, were attributable, at least in part, to sea lamprey control. Based on our analyses, predation by salmonines wasprimarily responsible for the reduction in alewife abundance during the 1970s and early 1980s. In turn, the decrease inalewife abundance likely contributed to recoveries of deepwater sculpin (Myoxocephalus thompsoni), yellow perch(Perca flavescens), and burbot populations during the 1970s and 1980s. Decrease in the abundance of all three domi-nant benthic macroinvertebrate groups, including Diporeia, oligochaetes, and sphaeriids, during the 1980s in nearshorewaters (50 m deep) of Lake Michigan, was attributable to a decrease in primary production linked to a decline inphosphorus loadings. Continued decrease in Diporeia abundance during the 1990s was associated with the zebra mussel(Dreissena polymorpha) invasion, but specific mechanisms for zebra mussels affecting Diporeiaabundance remain un-identified.

753Rsum : On trouvera ici une tude des changements dans le rseau alimentaire du lac Michigan de 1970 2000 etdes facteurs qui en sont responsables. Le contrle des populations de la grande lamproie marine (Petromyzon marinus)

et du gaspareau (Alosa pseudoharengus), qui a dbut dans les annes 1950 et 1960, a eu un impact considrable surle rseau alimentaire. Le rtablissement des populations du grand corgone ( Coregonus clupeaformis) et de la lotte(Lota lota) et laccroissement des populations de salmonids sont dus, au moins en partie, au contrle de la lamproie.

Can. J. Fish Aquat. Sci. 59: 736753(2002) DOI: 10.1139/F02-044 2002 NRC Canada

736

Received 22 March 2001. Accepted 25 January 2002. Published on the NRC Research Press Web site at http://cjfas.nrc.ca on24 May 2002.J16279

C.P. Madenjian1 and G.W. Fleischer. U.S. Geological Survey, Great Lakes Science Center, 1451 Green Road, Ann Arbor,MI 48105, U.S.A.G.L. Fahnenstiel. National Oceanic and Atmospheric Administration, Great Lakes Environmental, Research Laboratory, MuskegonField Station, 1431 Beach Street, Muskegon, MI 49441, U.S.A.T.H. Johengen. Cooperative Institute for Limnology and Ecosystem Research, University of Michigan, 2200 Bonisteel Boulevard,Ann Arbor, MI 48109, U.S.A.

T.F. Nalepa and H.A. Vanderploeg. National Oceanic and Atmospheric Administration, Great Lakes Environmental, ResearchLaboratory, 2205 Commonwealth Boulevard, Ann Arbor, MI 48105, U.S.A.P.J. Schneeberger. Michigan Department of Natural Resources, Marquette Fisheries Station, 484 Cherry Creek Road, Marquette,MI 49855, U.S.A.D.M. Benjamin. Atlantic States Marine Fisheries Commission, 444 Eye Street, Washington, DC 20005, U.S.A.E.B. Smith and J.R. Bence. Department of Fisheries and Wildlife, Michigan State University, East Lansing, MI 48824, U.S.A.E.S. Rutherford. Institute for Fisheries Research, University of Michigan, Ann Arbor, MI 48109, U.S.A.D.S. Lavis. U.S. Fish and Wildlife Service, Ludington Biological Station, 229 South Jebavy Drive, Ludington, MI 49431, U.S.A.D.M. Robertson. U.S. Geological Survey, Water Resources Division, 8505 Research Way, Middleton, WI 53562, U.S.A.D.J. Jude. Center for Great Lakes and Aquatic Sciences, University of Michigan, 501 East University, 717 Dennison Building,Ann Arbor, MI 48109, U.S.A.M.P. Ebener. Inter-Tribal Fisheries and Assessment Program, Chippewa Ottawa Resource Authority, 179 West Three Mile Road,Sault Ste. Marie, MI 49783, U.S.A.

1Corresponding author (e-mail: [email protected]).


2/18

Daprs nos analyses, la rduction de la densit des gaspareaux dans les annes 1970 et au dbut des annes 1980 estdue principalement la prdation par les salmonids. son tour, la rduction des gaspareaux a sans doute contribuau rtablissement des populations du chabot de profondeur (Myoxocephalus thompsoni), de la perchaude (Perca flaves-cens) et de la lotte dans les annes 1970 et 1980. Durant les annes 1980, la dcroissance de la densit des trois grou-pes dominants de macroinvertbrs benthiques, Diporeia, les oligochtes et les sphriids, dans les eaux ctires(50 m de profondeur) du lac Michigan est attribuable une diminution de la production primaire li une rductiondes apports de phosphore. La poursuite du dclin des densits de Diporeiadans les annes 1990 est associe

linvasion de la moule zbre (Dreissena polymorpha), bien que les mcanismes prcis de leffet des moules sur Dipo-reianaient pas encore t identifis.

[Traduit par la Rdaction]

Madenjian et al.Introduction

The objective of the first symposium on Salmonid Com-munities in Oligotrophic Lakes (SCOL-1) was to identify theeffects of the three major anthropogenic stressors of culturaleutrophication, fishery exploitation, and exotic species intro-

ductions on fish communities, especially salmonid commu-nities, in the five Laurentian Great Lakes (Wells and McLain1972, 1973). As part of the SCOL-1 effort, Wells andMcLain (1973) examined the changes in Lake Michigan fishcommunities from 1880 to 1970. One of the major conclu-sions from their study was that sea lamprey ( Petromyzonmarinus) and alewife (Alosa pseudoharengus) invasions be-tween 1930 and 1950 more profoundly affected fish commu-nities in Lake Michigan than either acceleratedeutrophication or fishery exploitation. Nevertheless, Wellsand McLain (1973) ascribed the population collapse oflarger deepwater ciscoes, including Coregonus nigripinnisand Coregonus johannae, to overfishing. Wells and McLain(1973) did acknowledge that accelerated eutrophication and

other pollution may have caused some changes in fish com-munities of Lake Michigan during 18801970; however, evi-dence to support this contention was not very strong. Inretrospect, SCOL-1 was a milestone in that it marked thefirst attempt to explain changes in fish communities of theLaurentian Great Lakes during 18801970 in the context ofanthropogenic stressors to the ecosystem.

The Lake Michigan food web has undergone tremendouschanges since 1970. For example, alewife abundance de-clined substantially during the 1970s and early 1980s (Judeand Tesar 1985; Eck and Wells 1987). Bloater (Coregonushoyi) abundance increased dramatically during the early1980s, peaked in the late 1980s, and declined during the1990s (Eck and Wells 1987; TeWinkel et al. 2002). Biomass

of the salmonine population increased steadily from 1965 to1986 (Stewart and Ibarra 1991). Phosphorus loadings toLake Michigan decreased during the 1980s as a result offederally mandated water quality control measures(Johengen et al. 1994). The three major groups of benthicmacroinvertebrates, namely the amphipod Diporeia,oligochaete worms (Oligochaeta), and fingernail clams(Sphaeriidae), decreased substantially in abundance in south-ern Lake Michigan at sites shallower than 50 m between1980 and 1993 (Nalepa et al. 1998). The midsummerepilimnetic phytoplankton community in offshore waters(>40 m deep) shifted from a blue-green and green algaedominated community in the 1970s to a phytoflagellate

dominated community in the 1980s (Fahnenstiel and Scavia1987a). The summer zooplankton community in offshorewaters changed abruptly during the early 1980s from a com-munity dominated by calanoid copepods to a communitydominated by the large-bodied cladoceran Daphniapulicaria (Evans and Jude 1986). The predatory cladoceranBythotrephes cederstroemi, an invader from Eurasia, became

established in Lake Michigan by 1986 (Lehman 1987). By1993, the zebra mussel (Dreissena polymorpha), an invaderfrom the Ponto-Caspian region, had colonized hard sub-strates within 7 km of shore between Waukegan and Michi-gan City (Nalepa et al. 1998). Presence of round goby(Neogobius melanostomus), a fish invader from the Ponto-Caspian region, was first documented in 1994 (Clapp et al.2001).

Herein, we document dynamics of the Lake Michiganfood web from 1970 through 2000, and then identify factorslikely responsible for these ecosystem changes. As inSCOL-1, we address the potential effects of anthropogenicstressors on fish communities in Lake Michigan. Since thetime of SCOL-1, the list of anthropogenic stressors to theGreat Lakes ecosystems has been expanded to include con-taminants and global climate change. We also re-examineconclusions from the SCOL-1 study, in light of the wealth ofinformation on the Lake Michigan food web that has be-come available since 1970. As has been surmised from stud-ies of other ecosystems (Likens 1989), we would expect thatlong-term observation and study are needed before we canbegin to understand the dynamics of an ecosystem as largeand complex as Lake Michigan. Therefore, we analyze theavailable long-term series of observations on phytoplankton,zooplankton, benthic macroinvertebrate, and fish communi-ties in Lake Michigan to identify factors that may be respon-sible for changes that have occurred since 1970. Aquaticecosystem structure can be regulated from the bottom up(McQueen et al. 1989), as well as from the top down(Kitchell and Carpenter 1993). How extensively have top-down and bottom-up effects pervaded the Lake Michiganfood web? With this synthesis, we answer that question.

Long-term trends in phosphorus loading

and nutrient dynamics

In Lake Michigan, phosphorus is the major nutrient thatlimits phytoplankton growth and has the greatest potential toadversely affect the lakes ecosystem (Schelske andStoermer 1971; Schelske et al. 1974). Fortunately, phospho-

2002 NRC Canada

Madenjian et al. 737


3/18

rus is also the easiest nutrient to control, primarily becausephosphorus has no gaseous phase. Thus, regulation of phos-phorus has been the primary management strategy for con-

trolling eutrophication in the Great Lakes. Billions of dollarshave been spent to reduce phosphorus inputs to the GreatLakes, and success has been documented in the lower GreatLakes (Johengen et al. 1994).

Phosphorus control programs substantially reduced phos-phorus loadings to Lake Michigan (Fig. 1). Total annualloading decreased from over 6.5 kilotonnes (kt) in 1980 toabout 4 kt in 1981, coincident with control of most pointsources from major water treatment plants and phosphorusbans on detergents. Loadings further decreased between1981 and the late 1980s, before appearing to level off duringthe 1990s (Fig. 1). During 1994 and 1995, annual load esti-mates ranged from 3 to 3.5 kt.

Based on available data (Johengen et al. 1994;

Makarewicz et al. 1998; T. Johengen, unpublished data), ni-trate and silica concentrations in epilimnetic waters of LakeMichigan increased between 1976 and 1999. Increases wereapparent for both spring maximum and summer minimumconcentrations within the epilimnion. Increased nitrate levelshave likely resulted from increased tributary loads, in re-sponse to human population growth and increases in indus-trial and agricultural activity (Vollenweider 1970), as well asfrom increased atmospheric loading, as was clearly demon-strated for Lake Superior (Bennet 1986). Silica concentra-tions would be expected to increase from an increase inanthropogenic loading, a decrease in the burial rate of silicato the sediment, or a decrease in rates with which dissolvedsilica is biologically utilized. Rate of use of silica by biota

would decrease if species composition of the diatom com-munity shifted toward smaller or less-silicified diatoms, or ifthere was an overall decrease in production rates of the dia-tom community.

Dynamics of phytoplankton and primary

production

In toto, the evidence indicated that primary production inoffshore waters of Lake Michigan trended neither upwardnor downward during the last 25 years. Total phytoplanktonabundance, measured as chlorophylla concentration, in off-

shore waters (>50 m deep) of Lake Michigan varied withouttrend during 19831998 (Fig. 2a), a finding corroborated byFahnenstiel et al. (1998), who showed that phytoplanktonabundance, measured as carbon, in offshore waters of LakeMichigan did not change between the 19861988 and 19931995 time periods. Addition of chlorophyll data fromBrooks and Torke (1977) to the time series demonstrated notrends in offshore phytoplankton abundance since 1973

(Fig 2a). Similarly, total phosphorus concentration (TP) inthe water column during spring, an indicator of primary pro-duction (Fahnenstiel et al. 1998), fluctuated about a constantlevel during 19761998 (Fig. 2b).

In contrast, species composition of the phytoplanktoncommunity in offshore waters of Lake Michigan underwentsubstantial changes in the past 30 years. Community struc-ture during summer stratification shifted from one domi-nated by blue-green and green algae during the early 1970sto one dominated by phytoflagellates during the early 1980s(Fahnenstiel and Scavia 1987a). Summertime dominance ofphytoflagellates has continued into the 1990s (Makarewicz

2002 NRC Canada

738 Can. J. Fish Aquat. Sci. Vol. 59, 2002

Fig. 1. Annual estimates of total phosphorus loading to LakeMichigan, including Green Bay, for 19741991 and 19941995.

Fig. 2. (a) Average annual chlorophyll aconcentration and(b) average spring total phosphorus concentration (from Marchuntil water temperature reached 4C) in Lake Michigan at off-shore (100110 m) stations off Grand Haven (!) and Muskegon(), Mich., 19761998. Also included are chlorophyll adata foroffshore Lake Michigan waters off Milwaukee, Wis., 19731974() (Brooks and Torke 1977).


4/18

et al. 1998). Diatoms have composed the bulk of thephytoplankton community during the spring isothermal mix-ing period since 1970.

In Lake Michigan, a deep chlorophyll layer typicallyforms in the upper hypolimnion and lower metalimnionduring summer stratification, and this layer contributesbetween 5 and 10% of the annual total chlorophyll and an-

nual primary production within the lake (Fahnenstiel andScavia 1987b). Size of the deep chlorophyll layer in-creased between the 1970s and 1980s. This increase wasattributed to greater light penetration caused by increasedgrazing due to increased abundance of the very large-bodied zooplankter D. pulicaria(Scavia et al. 1986). Thischange in epilimnetic zooplankton composition did notpersist for more than a few years (Evans 1992), and a shiftback to smaller-bodied zooplankton apparently led to a re-duction in the size and concentration of the deep chloro-phyll layer (Fahnenstiel and Scavia 1987c; G. Fahnenstiel,unpublished data).

Although a long-term series of observations onphytoplankton abundance was not available for most of the19702000 time period, the available data suggested thatphytoplankton abundance and production in nearshore wa-ters of Lake Michigan have decreased during the past30 years. Schelske et al. (1980) intensively surveyed springTP and chlorophyll a concentrations in nearshore waters ofthe eastern side of the lake during 1972. Nearshore watersin the vicinity of the Cook Power Plant in the southeasternsection of the lake was consistently sampled for chloro-phyll a concentration during 19731981. Similar samplingwas conducted during the recent Episodic Events GreatLakes Experiment (EEGLE) study from 19982000 in thesame nearshore waters of Lake Michigan (G. Fahnenstiel,unpublished data). Comparison of data between the twotime periods indicated that chlorophyll a concentration de-creased by more than 50% and spring TP decreased by40% between the 1970s and the 19982000 time period.These declines in phytoplankton abundance and productionwere likely attributable to the reduction in phosphorusloading (Fig. 1).

Owing to lack of regular assessments of the species com-position of the phytoplankton community in nearshore wa-ters, changes in phytoplankton community structure weredifficult to firmly establish. Diatoms were the predominantgroup during spring mixing in the 1990s (G. Fahnenstiel, un-published data), just as they were in the 1970s (Chang andRossman 1988). The springtime community was dominatedby small centric diatoms (such as Stephanodiscus) in

nearshore waters and by large centric diatoms (such asAulacoseira) in offshore waters (Fahnenstiel and Scavia1987a; G. Fahnenstiel, unpublished data). Relative abun-dance of blue-green algae during summer increased from the1970s to the 1980s in nearshore waters (Chang and Rossman1988). Blue-green algae continued to be important innearshore waters during summer stratification in the 1990s(G. Fahnenstiel, unpublished data).

During the 1970s, silica may have limited the productionof diatoms during the end of the spring diatom bloom inboth nearshore and offshore waters of Lake Michigan(Schelske et al. 1986; Scavia and Fahnenstiel 1987).

Dynamics of major benthic

macroinvertebrate groups

The three major benthic macroinvertebrate groups foundin Lake Michigan have been the amphipod Diporeia,oligochaete worms (Oligochaeta), and fingernail clams(Sphaeriidae) (Nalepa et al. 1998). Benthic surveys con-

ducted in the southern basin have been sufficiently consis-tent and thorough to construct a long-term record ofabundances of these groups since 1965. Overall, abundancesofDiporeia, oligochaetes, and sphaeriids increased betweenthe 1960s and the early 1980s (Fig. 3). The magnitude of theincrease varied with group and depth interval. Long-termtrends in abundance during 19801993 differed betweennearshore (50 m) and offshore (>50 m) depth zones ofsouthern Lake Michigan (Nalepa et al. 1998, 2000a). In thenearshore, abundances of all three groups declined between1980 and 1993 (Fig. 3a, 3b). In contrast, abundances ofDiporeia, oligochaetes, and sphaeriids in the offshore zoneexhibited no distinct trends over the same time period(Fig. 3c, 3d). This difference in temporal trends of benthic

macroinvertebrate abundances between nearshore and off-shore zones may have been due to differences in temporaltrends in phytoplankton abundance between the two zones.In large deep lakes like Lake Michigan, abundance of thedominant groups of benthic macroinvertebrates tends to bedirectly proportional to the amount of available food(Saether 1980). Increased phytoplankton abundance leads toincreased amounts of organic material settling to the bottomregion, thereby providing more potential food formacrobenthos. As noted above, phytoplankton abundanceand primary production appeared to decrease in nearshorewaters of Lake Michigan between the 1970s and 1990s,whereas no change was detected in offshore waters. The de-crease in primary production in nearshore waters was proba-bly due to the reduction in phosphorus loadings to the lake.Nutrient inputs from tributaries is better coupled with pri-mary production in nearshore waters than in offshore waters(Schelske 1977). Finally, the zebra mussel invasion, in addi-tion to the reduction in primary production, may have partlycontributed to the decrease in Diporeia abundance in thenearshore region during the early 1990s (Nalepa et al.2000a).

Continued decline of Diporeia in the nearshore regionduring the 1990s coincided with the invasion of Lake Michi-gan by the zebra mussel (Nalepa et al. 2000a, 2000b). By1993, zebra mussels had colonized hard substrates within7 km of shore between Waukegan and Michigan City. Zebra

mussel density continued to increase at sites


5/18

in Lake Michigan, and Diporeiaexhibits a higher lipid con-tent than other benthic macroinvertebrates (Gardner et al.1985). If Diporeia were sufficiently reduced in abundanceon a lake-wide basis, then fish heavily reliant on Diporeiaasa food source would have to switch to other prey. Conse-quences of such diet switching on the growth and survival ofthese fishes remain unknown.

Although surveys by Nalepa (1987) and Nalepa et al.(1998) have documented changes in macroinvertebrate com-munities in areas deeper than 15 m, community changes

have also occurred in shallower regions of Lake Michigan.The zebra mussel invasion has resulted in increased abun-dance of chironomids and the amphipod Gammarus in theshallow, rocky habitats of southern Lake Michigan (Kuhnsand Berg 1999).

Dynamics of zooplankton

Long-term trends in zooplankton community structure andabundance have been documented for nearshore waters(50 m deep) of southeastern Lake Michigan in the vicinityof the Cook Power Plant between 1972 and 1984 by Evans

and Jude (1986) and Evans (1992), and for a large portion ofthe offshore region (>50 m deep) during 19831992 byMakarewicz et al. (1995). In deeper (30- to 40-m) waterssampled by Evans and Jude (1986), summer daphnid assem-blage shifted from a 19721976 community strongly domi-nated by the medium-bodied Daphnia retrocurvato a 19771981 community strongly dominated by the large-bodiedD. galeata mendotae, to a 19821984 community dominatedby the very large-bodied D. pulicaria. In contrast, Daphniaretrocurva dominated the daphnid community in shallower

waters in the Cook Plant vicinity from 1972 to 1984. Addi-tionally, abundance of daphnids in shallower waters wassubstantially higher during 19721980 than during 19811984. Evans and Jude (1986) attributed this decrease indaphnid abundance to increased predation pressure exertedby the yellow perch (Perca flavescens) population, which in-creased dramatically in size between 1980 and 1984. Evansand Jude (1986) concluded that the appearance of the verylarge-bodied D. pulicariain the deeper waters during 19821984 was possibly related to a decline in alewife abundanceduring the 1970s and early 1980s. Wells (1970) contendedthat increased alewife abundance in offshore waters of Lake

2002 NRC Canada


Fig. 3. Average annual abundances of the dominant macroinvertebrate groups in the southern basin of Lake Michigan, in selected yearsover 19651993, within (a) the 16- to 30-m depth contour, ( b) the 31- to 50-m depth contour, ( c) the 51- to 80-m depth contour, and(d) the >80-m depth contour. Error bar represents one standard error for average total annual abundance. Solid bar represents Diporeia,open bar represents Sphaeriidae, and hatched bar represents Oligochaeta.


6/18

Michigan during the late 1950s and early 1960s led to a shiftin the size structure of the summer zooplankton communityfrom large-bodied to small-bodied copepods andcladocerans. Evidence to support the contention that highabundance of alewives in 1966 resulted in a zooplanktoncommunity dominated by small-bodied forms was compel-ling; however, relatively low alewife abundance did not nec-essarily imply that very large-bodied zooplankters would

dominate the zooplankton community (Evans 1992).Although species composition of the crustacean zooplank-ton community changed considerably during 19831992 inoffshore waters of Lake Michigan, crustacean zooplanktonbiomass varied without trend during that time (Makarewiczet al. 1995). Furthermore, in comparing their data with ob-servations by Scavia et al. (1986), Makarewicz et al. (1995)concluded that crustacean zooplankton biomass during thesummer in offshore waters of Lake Michigan showed notrend between 1975 and 1992. Daphnia pulicariawas an im-portant constituent of the daphnid community in 1983, butsubsequently declined in abundance until nearly disappear-

ing from the lake by 1987 (Makarewicz et al. 1995). Over-all, Diaptomus dominated the zooplankton community dur-ing 19831992. Increases in rotifer biomass between 1983and 1992 were correlated with increases in unicellular algae(Makarewicz et al. 1995). Based on correlation analysis,Makarewicz et al. (1995) suggested that alewives during19831992 were probably not playing a dominant role in af-

fecting zooplankton biomass and size structure except forthe size structure of calanoid copepods. However, alewifeabundance remained relatively low during this time period(Fig. 4). Preliminary examination of zooplankton samplesduring 19941999 suggested that no major changes in spe-cies composition or abundance had occurred since theMakarewicz et al. (1995) study (J. Cavaletto, NOAA GreatLakes Environmental Research Laboratory, Ann Arbor, MI48105, U.S.A., personal communication).

Invasion of Lake Michigan by the predatory cladoceranBythotrephes cederstroemiin 1986 appeared to have causedsome changes in zooplankton community structure(Makarewicz et al. 1995). For example, the decrease inDaphnia retrocurva abundance after 1986 was most likely

attributable to predation by Bythotrephes (Lehman andCceres 1993).Other than the increase in rotifer biomass between 1983

and 1992, evidence for changes in zooplankton communitystructure and abundance originating from bottom-up effectswas not strong (Makarewicz et al. 1995). Furthermore, theissue of whether changes in top-down control from alewivesor from Bythotrephesled to changes in phytoplankton abun-dance or composition remains contentious (Evans 1992;Makarewicz et al. 1998).

Abundance of Mysis relicta, a common prey of alewife,bloater, rainbow smelt (Osmerus mordax), and lake whitefish(Coregonus clupeaformis), in southern Lake Michigan hasremained relatively stable from the 1970s through 1998

(Pothoven et al. 2000). Mysis is an important predator ofzooplankton in Lake Michigan. Whereas micrograzers (pro-tozoans and ciliates) are an important component of theLake Michigan food web (Fahnenstiel et al. 1998), insuffi-cient data are available to characterize temporal trends intheir abundance or their community structure.

Dynamics of the prey fish community

AlewifeThe alewife invasion during the 1940s greatly altered the

fish community of Lake Michigan. Declines in abundancesof emerald shiner (Notropis atherinoides), deepwater sculpin

(Myoxocephalus thompsoni), yellow perch, and bloater dur-ing the 1960s have been attributed to the alewife invasion(Wells and McLain 1973). Also, alewife predation on laketrout (Salvelinus namaycush) fry may represent a serious im-pediment to lake trout rehabilitation in the lower four GreatLakes (Krueger et al. 1995).

One of the reasons for launching a major stocking pro-gram for salmon and trout in Lake Michigan in 1965 was toreduce alewife abundance, which had reached nuisance lev-els during the 1960s (Hatch et al. 1981; Rutherford 1997).Alewives have dominated the diet of these stockedsalmonines (Stewart and Ibarra 1991; Madenjian et al.

2002 NRC Canada


Fig. 4. (a) Estimated lake-wide biomass of alewife (light shad-ing), rainbow smelt (solid), deepwater sculpin (open), and bloater(dark shading) in Lake Michigan, 19731999. Estimates basedon annual bottom trawl surveys performed by the USGSGLSC.(b) Estimated lake-wide biomass of lake trout (light shading),brown trout (dark shading), rainbow trout (open), coho salmon(vertical bars), and chinook salmon (solid) in Lake Michigan,

19651998. Biomass was estimated using age-specific populationmodels. See text for more details on modeling.


7/18

1998a), and to a lesser extent, have also been eaten byburbot (Fratt et al. 1997).

Stocking of salmonines appeared to be effective in con-trolling alewife abundance. Alewife abundance decreasedduring the 1970s and early 1980s, coincident with thebuildup of salmonines (Fig. 4). Alewife abundance has nei-ther trended upward nor downward since the early 1980s,

concomitant with relatively high levels of salmonine bio-mass within the lake. Lake-wide alewife biomass in 1973,based on bottom trawl surveys conducted by the U.S. Geo-logical Survey (USGS) Great Lakes Science Center (GLSC),was estimated at 125 kt, whereas lake-wide biomass esti-mates averaged only 37 kt between 1984 and 1999 (Fig. 4a).Furthermore, additional GLSC trawling data indicated thatthe alewife decline was well underway by 1970 (Krause1999). Although the lake-wide surveys of prey fish popula-tions did not begin until 1973, GLSC bottom trawling wasconducted at selected ports along the eastern and southernportions of the lake since 1962 (Hatch et al. 1981). Krause(1999), accounting for the reduced set of sampling locationsduring 19621972 by using a mixed model approach, esti-

mated that lake-wide abundance of alewives in Lake Michi-gan was about twice as high in 1969 than it was in 1973.Further, the 1966 abundance was estimated at more than fivetimes higher than the 1969 abundance. However, the de-crease between 1966 and 1969 was partially attributable to adie-off occurring in 1967 (Brown 1972). Finally, Madenjianet al. (1986), using trend analysis, demonstrated a significantdecline in alewife abundance in southeastern waters of LakeMichigan between 1973 and 1982.

Eck and Brown (1985) proposed that cold temperaturesregulated alewife abundance in Lake Michigan, because aportion of the decline in alewife abundance during the 1970sand early 1980s coincided with relatively cold weather.However, 19731999 data for alewife abundance and tem-perature do not support this contention. The alewife popula-tion did not rebound even though the cold years of 19771982 were followed by relatively warm years of 1983 and19851992 (Fig. 4a; T. Karl, NOAA National Climatic DataCenter, Asheville, NC 28801, U.S.A., unpublished data).Furthermore, annual alewife abundance during the 19731999 time period was poorly correlated (r = 0.03, P =0.8711, n = 27) with heating degree days at Muskegon(Mich.), an indicator of winter severity (Eck and Brown1985). In contrast, annual alewife abundance during the19731998 time period showed a significant, negative rela-tionship (r= 0.54, P= 0.0058, n = 26) with estimated an-nual lake-wide consumption of alewives by salmonines.

The lengthfrequency distribution of Lake Michigan ale-wives did not drastically change between 1973 and 1999.Alewife growth rate and condition did not exhibit a trend be-tween 1984 and 1995; however, alewife condition declinedby about 15% after 1995 (C. Madenjian, unpublished data).Nearly all of the diet of adult alewives in Lake Michiganfrom 1970 to 1995 has been composed of zooplankton(chiefly cladocerans and copepods), Diporeia, and Mysis(Davis et al. 1997). Lipid content in alewives neither in-creased nor decreased between 1969 and 1995 (Madenjian etal. 2000), and modal length of adult alewives in bottomtrawls has shown no trend since 1973 (C. Madenjian, unpub-

lished data). The recent decline in alewife condition may belinked to the recent decline in Diporeiaabundance.

BloaterAbundance of age-1 and older (hereafter referred to as

adult) bloater in Lake Michigan has undergone dramaticchanges in the last 30 years (Fig. 5a). Bloater recruitmentstarted to increase in 1977, and relatively large year-classeswere produced during 19801990 (Fig. 5b). Since 1992,however, year-class strength has been very weak. Temporaltrends in abundance of adult bloater corresponded well with

temporal trends in age-0 abundance. Adult abundance in-creased exponentially during 19771983 and then continuedto increase, albeit at a slower rate, until 1989 (Fig. 5 a).Since peaking in 1989, adult abundance has decreased. Ex-amination of bottom trawl data for years 19621972, whenspatial coverage of the lake during fall surveys was incom-plete, suggested that adult bloater abundance had alsopeaked sometime during the late 1950s or early 1960s (G.Fleischer, unpublished data).

Reasons for rise and fall in bloater abundance between1970 and 2000 were not clear. Decline in bloater abundanceduring the 1960s was attributed to alewives interfering with

2002 NRC Canada


Fig. 5. Abundance of (a) adult (>120 mm total length) and(b) age-0 bloaters in Lake Michigan, 19731999, based on an-nual bottom trawl surveys performed by the USGSGLSC. Catchper unit of effort (CPUE) expressed in number of fish caught per10-min bottom trawl tow. Horizontal line represents mean catchrate for entire time period.


8/18

bloater reproduction (Wells and McLain 1973). However,alewife abundance was relatively low during 19921999,when bloater year-class strength was very low. Jude andTesar (1985) proposed that release from commercial exploi-tation, more than release from alewife interference, stimu-lated bloater recovery beginning in the late 1970s. However,exploitation rate on bloaters during the late 1980s and early

1990s was only about 1% per year, yet bloater abundancehas been declining since 1989. We propose a new hypothe -sis to explain the pattern in bloater abundance in Lake Mich-igan for the past 40 years: bloater abundance in LakeMichigan follows a cycle with a period of about 30 years.Further, the cycle appears to be mainly independent of inter-actions with other fish populations or humans, and sometype of density-dependent feedback may be involved inmaintaining these cycles.

Growth and lipid content of bloaters in Lake Michiganhave varied with population size, whereas diet compositionhas been consistent. As abundance increased during the1980s, both growth rate and lipid content decreased(Madenjian et al. 2000; TeWinkel et al. 2002). DiporeiaandMysis have remained the mainstay of adult bloater diet inLake Michigan (Davis et al. 1997). Sex ratio of the bloaterpopulation became skewed, with females predominating,during the late 1980s (TeWinkel et al. 2002).

Rainbow smelt

Rainbow smelt became established in Lake Michigan by1936 (Wells and McLain 1972, 1973). This naturalized pop-ulation of rainbow smelt continues to be exploited by com-mercial fisheries, albeit at lesser yields compared withbloater (S. Nelson, USGS Great Lakes Science CenterCOMCAT database, Ann Arbor, MI 48105, U.S.A.). Adultrainbow smelt have remained an important diet constituent

for intermediate-sized (400- to 600-mm total length) laketrout in nearshore waters of Lake Michigan since the 1970s(Madenjian et al. 1998a). Also, rainbow smelt representedan important diet component for salmon during 19831987(Stewart and Ibarra 1991).

While relatively low during 19731977, lake-wide bio-mass (as estimated via bottom trawling) of rainbow smelt in-creased during 19771982 (Fig. 4a). On average, lake-widebiomass of rainbow smelt was relatively high during 19821992, but then declined between 1992 and 1997. Reasons forthe decline were not obvious. It seems unlikely that preda-tion on rainbow smelt by salmonines has increased substan-tially since 1987 (see section on Dynamics of the salmonine

community).Mean total length of adult rainbow smelt in Lake Michi-gan decreased approximately 10% from 1973 to 1987, withno appreciable decrease since 1987 (G. Fleischer, unpub-lished data). The decrease in rainbow smelt size was mostlikely attributable to an increase in predation due to buildupof salmonine biomass. Lipid content of rainbow smelt didnot exhibit any definitive trend from the early 1970s to 1995(Madenjian et al. 2000). Diet of adult rainbow smelt in LakeMichigan has consisted primarily of Mysis and small fish,especially young-of-the-year (YOY) alewife, during both the1970s and 19941995 (Davis et al. 1997).

Deepwater sculpin

A substantial portion of the diet of burbot in Lake Michi-gan is composed of deepwater sculpins (Fratt et al. 1997).To a lesser degree, deepwater sculpins are also eaten by laketrout (Madenjian et al. 1998a). Abundance of deepwatersculpin in Lake Michigan increased rapidly between 1973and 1983 (Fig. 6). After peaking in the early 1980s, deep-water sculpin lake-wide biomass declined until 1991, butshowed no further decline after 1991 (Fig. 4a). Rapid in-crease in population size during 19731983 was most likely

attributable to the decrease in alewife abundance during thattime; Wells and McLain (1972, 1973) proposed that the de-cline in deepwater sculpin abundance during the 1960s wasdue to interference, by alewives, with deepwater sculpin re-cruitment. Alewives may consume the pelagic fry of deep-water sculpins. The decrease in deepwater sculpinabundance between the early 1980s and 1991 was probablydue to increased predation by the recovering burbot popula-tion, which increased in size during the 1980s but appearedto level off in size during the 1990s (Fig. 7). Diporeia hasdominated deepwater sculpin diet in Lake Michigan during19701995 (Davis et al. 1997).

2002 NRC Canada


Fig. 6. Abundance of slimy (!) and deepwater () sculpins inLake Michigan, 19731999, based on annual bottom trawl sur-veys performed by the USGSGLSC. Catch per unit of effort(CPUE) expressed in (a) number and (b) kilograms of fishcaught per 10-min bottom trawl tow.


9/18

Slimy sculpinSlimy sculpin (Cottus cognatus) has typically been preyed

upon by juvenile lake trout in Lake Michigan, at least sincethe 1970s (Madenjian et al. 1998a). To a lesser degree, slimysculpins are also eaten by burbot (Fratt et al. 1997). Long-term trends in slimy sculpin abundance, as determined bybottom trawl surveys, appeared to be related to changes inthe spatial distribution of newly stocked lake trout. Slimysculpin abundance decreased during the 1970s and early1980s, remained relatively low during most of the 1980s,and then increased during the 1990s (Fig. 6). Prior to 1986,most of the lake trout stocking in Lake Michigan occurredrelatively close to shore, but stocking emphasis was placed

on offshore refuge areas beginning in 1986 (Holey et al.1995). An increase in predation on slimy sculpins with in-creased lake trout population size during the 1970s probablycaused the decrease in slimy sculpin abundance during the1970s (Eck and Wells 1987). With the shift to predominantlyoffshore stocking in 1986, predation on slimy sculpins in ar-eas of the lake trawled by USGSGLSC likely declined, andthis decline in predation allowed the increase in slimy scul-pin abundance observed during the 1990s. Slimy sculpins inLake Michigan have fed predominantly on Diporeia sincethe 1960s (Davis et al. 1997).

Dynamics of the salmonine community

BackgroundStocking of salmonine fishes into Lake Michigan has had

a profound effect on the lakes food web. Large-scale stock-ing of salmonines began in 1965, and has continued to thepresent. The rationale for stocking salmonines was (i) tocontrol abundance of alewives, which had become a nui-sance, (ii) to initiate a valuable sport fishery, (iii) to possiblyinitiate naturalized populations of introduced salmonines,and (iv) to rehabilitate the lake trout population (Holey et al.1995; Rutherford 1997). The five species of salmonines cur-rently stocked in Lake Michigan are lake trout, chinooksalmon (Oncorhynchus tshawytscha), coho salmon

(Oncorhynchus kisutch), rainbow trout (Oncorhynchusmykiss), and brown trout (Salmo trutta).

Long-term trends in biomass, production, andconsumption

Lake-wide biomass and production of the five salmoninespecies was estimated using age-specific population modelsto track abundance at age over time. For all species exceptchinook salmon, our modeling approach was similar to theone used by Jones et al. (1993) for Lake Ontario salmonines.Using estimates of gross conversion efficiency (GCE) frombioenergetics modeling (Stewart and Ibarra 1991), lake-wideconsumption of prey by salmonine populations was calcu-lated using the productionconversion efficiency method ofNey (1990) where gross production is divided by GCE to es-timate consumption. For chinook salmon, population model-ing techniques outlined in Benjamin and Bence (2002) wereused to estimate biomass, production, and consumption. Re-cruitment was quantified as the number of individuals enter-ing the lake population, and equaled the sum of hatchery andnaturally reproduced production. Salmonine stocking recordshave been summarized by the U.S. Fish and Wildlife Service(C. Bronte, Green Bay Fishery Resources Office, Green Bay,WI 54311, U.S.A., salmonine stocking database) andBenjamin (1998). Natural reproduction by chinook and cohosalmon, as well as rainbow trout, was quantified using previ-ously developed methods (Rutherford 1997). We assumedthat natural recruitment to lake trout and brown trout popula-tions has been negligible since 1965 (Holey et al. 1995;Rutherford 1997).

Salmonine biomass in Lake Michigan increased from0.039 kt in 1965 to nearly 27 kt in 1986, then declined dur-ing the late 1980s, driven by a decline in abundance of chi-nook salmon, which has accounted for over 40% of the total

salmonine biomass since 1980 (Fig. 4b). This decrease inchinook salmon biomass has been attributed to an outbreakof bacterial kidney disease (BKD) within the chinooksalmon population (Kabre 1993). Presumably due to a de-crease in BKD-induced mortality, chinook salmon biomassincreased during 19941998, and salmonine biomass was es-timated at nearly 30 kt in 1998 (Fig. 4b). For the most part,temporal trends in gross production mimicked trends insalmonine biomass. Annual gross production of salmoninebiomass increased from 0.235 kt in 1965 to nearly 23 kt in1998. Ratio of annual gross production to biomass (P/B)generally varied from 0.8 to 1.0.

Consumption of prey by Lake Michigan salmonines in-creased from only 1 kt in 1965 to 130.5 kt in 1998 (Fig. 8).

Since 1977, the chinook salmon population has been respon-sible for approximately 50% of the consumption bysalmonines (Fig. 8a). Most predation by salmonines in LakeMichigan has been directed at alewives (Fig. 8b). Alewives(both small and large sizes combined) represented, on aver-age, 69.2% of total annual consumption from 1985 to 1998,with a maximum of 79.7% in 1998. Annual consumption ofalewives by salmonines exceeded 90 kt in 1996 and 1998.Chinook salmon have been responsible for 62% of the pre-dation on alewives by salmonines since 1980.

An important factor driving salmonine biomass and pro-duction, as well as consumption of prey by salmonines, wastrends in the recruitment of young fish to the salmonine

2002 NRC Canada


Fig. 7. Abundance of burbot in Lake Michigan, 19731999,based on annual bottom trawl surveys performed by the USGSGLSC. Catch per unit of effort (CPUE) expressed in number offish caught per 10-min bottom trawl tow.


10/18

stocks. Natural recruitment of chinook salmon has increasedfrom 1970 to the present. We estimated that annual naturalrecruitment of chinook salmon approached 2 million smoltsduring the 1990s. In some years, natural recruitment repre-sented nearly 40% of the total annual recruitment of chinooksalmon (Rutherford 1997).

Long-term trends in diet and growth

Growth rate of chinook salmon in Lake Michigan de-clined during the early 1980s (Stewart and Ibarra 1991), butthen increased during the late 1980s and early 1990s (Wes-ley 1996). The increase could have been a compensatory re-sponse to the decline in chinook salmon abundancefollowing the BKD outbreak (Wesley 1996). Growth rateagain declined during the late 1990s. Alewife has alwaysbeen the most prominent component of chinook salmon diet.However, alewife was of greater importance in chinooksalmon diet during the 1970s than during 19831987, whenrainbow smelt was an important diet constituent and whenalewives in the diet tended to be smaller than in the 1970s or19941995 (Stewart and Ibarra 1991; R. Elliott, U.S. Fish

and Wildlife Service, Green Bay, WI 54311, U.S.A., per-sonal communication).

Growth rate of lake trout remained relatively constant be-tween the 1970s and 2000 (Madenjian et al. 1998a; J. Jonas,Michigan Department of Natural Resources, Charlevoix, MI49720, U.S.A., unpublished data). Overall, alewife has re-mained the most important component in adult lake trout

diet. Nevertheless, bloater was a noteworthy constituent,contributing up to 35% of the diet of adult lake trout innearshore waters of southeastern Lake Michigan between1985 and 1995. In contrast, adult lake trout diet in nearshoreWisconsin waters did not appear to change appreciably sincethe 1970s (Madenjian et al. 1998a). Lipid content of laketrout in Lake Michigan did not substantially change between1970 and 1995 (Madenjian et al. 2000).

Changing fisheriesSport fisheries for Lake Michigan salmonines greatly ex-

panded from the mid-1960s to 1985, in response to in-creased salmonine abundance (Bence and Smith 1999). Withdecline of chinook salmon abundance in the late 1980s, an-gling effort declined. Harvest of rainbow trout increased dur-ing the late 1980s and has remained relatively high duringthe 1990s (Bence and Smith 1999). Chinook salmon harvestnearly doubled between 1994 and 1998, but the 1998 harvestwas still less than 30% of the 1986 harvest. Lake trout havebeen harvested continually both recreationally and commer-cially since 1970 (Bence and Smith 1999).

Lake trout rehabilitationLake trout were extirpated from Lake Michigan during the

1950s, presumably because of overfishing and sea lampreypredation (Wells and McLain 1973). Since 1965, an averageof about 2.4 million lake trout yearlings have been stocked

into Lake Michigan each year. As part of the lake trout reha-bilitation effort, two offshore refuges were established dur-ing 19841985 (Holey et al. 1995). The Northern Refugeencompassed a complex of shallow-water (0- to 20-m)spawning reefs, whereas the Southern or Midlake Refuge in-cluded four large deep-water (45- to 80-m) spawning reefs.Fishing for lake trout continues to be prohibited in both ref-uges. Emphasis on stocking the two refuges began in 1986,when stocking rates in nearshore zones sharply declined.

Although the stocking program has yielded a substantialnumber of spawners, documentation of natural reproductionby lake trout in Lake Michigan since 1965 has been ex-tremely sparse (Holey et al. 1995). This negligible amountof successful natural reproduction appeared to be due to sev-

eral factors. Contaminants, once thought to strongly interferewith lake trout reproduction in Lake Michigan, have recentlybeen shown not to be a significant hindrance, while thiaminedeficiency is now seen as a likely factor (Fitzsimons et al.1999). By previously established criteria, density of spawn-ing lake trout in the Northern Refuge was sufficiently highto allow for successful natural reproduction (Madenjian andDeSorcie 1999). Perhaps hatchery reared lake trout are soinefficient at reproducing that considerably higher densitiesof spawners or a higher density of older (age >10 years)spawners are required for successful natural reproduction.Density of lake trout spawners on Sheboygan Reef in theMidlake Refuge exceeded the relatively high density of 175

2002 NRC Canada


Fig. 8. (a) Estimated consumption of prey by lake trout (lightshading), brown trout (dark shading), rainbow trout (open), cohosalmon (vertical bars), and chinook salmon (solid) in Lake Mich-igan, 19651998. (b) Estimated consumption of large alewives(solid), small alewives (open), and other fish (shading) bysalmonines in Lake Michigan, 19651998. Consumption was es-timated using age-specific population models coupled with

bioenergetics models. See text for more details on modeling.


11/18

fish per 305 m of gill net during the late 1990s and a sub-stantial proportion of these spawners was composed of fish>10 years old, yet no signs of natural recruitment have been

observed (M. Toneys and P. McKee, Wisconsin Departmentof Natural Resources, Sturgeon Bay, WI 54235, U.S.A., un-published data). Perhaps alewife predation on lake trout fryhas also been playing an important role in suppressing natu-ral reproduction by lake trout (Krueger et al. 1995).

Dynamics of other fish populations

Sea lampreyTreatment of Lake Michigan tributaries, beginning in the

late 1950s, with the lampricide 3-trifluoromethyl-4-nitrophenol (TFM) led to an estimated 8090% decline inthe abundance of adult sea lamprey in Lake Michigan by1966 (Lavis et al. 2002). Lampricide treatment in streams

continues to be the primary strategy for controlling the sealamprey population in Lake Michigan.Population sizes of spawning-phase sea lamprey in all

Lake Michigan tributaries each year since 1977 have beenestimated using a regression model, which incorporated rela-tive abundance data from assessment traps in tributaries andestimates of stream discharge rates (Lavis et al. 2002). Thislong-term series indicates that numbers of spawning sealampreys in Lake Michigan tributaries increased during19771984, showed no trend between 1984 and 1995, andthen increased again during 19951999. The two increasesoccurred primarily in northern tributaries to the lake. Thefirst increase was attributed to an immigration of adult sealampreys from northern Lake Huron, whereas the second in-

crease was probably due to a decrease in sea lamprey controlmeasures for Lake Michigan (G. Christie, Great Lakes Fish-ery Commission, Ann Arbor, MI 48105, U.S.A., personalcommunication). Lake-wide estimates of spawning sea lam-prey population size ranged between 32 000 and 117 000 sealampreys (Lavis et al. 2002).

Long-term trends for incidence of fresh sea lampreywounds on lake trout in Lake Michigan mimicked long-termtrends for sea lamprey abundance. Wounding rates in south-ern waters of the lake have remained nearly unchanged forthe last 25 years, whereas an increase in wounding rates hasbeen observed in northeastern waters since 1978 (Lavis et al.2002). Overall, lamprey-induced annual mortality, estimated

from wounding rate, for Lake Michigan lake trout was rela-

tively low, averaging less than 7% for most years during19841998.

Lake whitefishLake whitefish is the most important commercial fish in

Lake Michigan. Tribal and state-licensed commercial fishersfrom Michigan and Wisconsin continue to exploit lakewhitefish. Lake whitefish are taken in trap nets, gill nets,trawls, and pound nets.

Since 1970, annual commercial harvest of lake whitefishhas more than tripled (Fig. 9), reflecting a lake-wide in-crease in abundance. A dramatic decline in lake whitefishabundance during the late 1950s was linked primarily withsea lamprey predation (Wells and McLain 1972). Thus, there

can be little doubt that the impressive recovery of the lakewhitefish population in Lake Michigan since 1970 was par-tially driven by sea lamprey control (Eck and Wells 1987).Stocking of lake trout also afforded extra protection fromsea lamprey predation because lake trout are favored preyfor sea lampreys (Wells and McLain 1972). The decrease inmean size of adult rainbow smelt during the 1970s and1980s and decreased rainbow smelt abundance during the1990s may have also contributed to the lake whitefish recov-ery because large rainbow smelt may prey upon lakewhitefish fry (Wells and McLain 1973). During the last 30years, exploitation has not impaired the ability of the adultlake whitefish population to produce large year-classes. Thisis true even though contemporary stocks of lake whitefish

have been subjected to high exploitation rates, leading to to-tal annual mortality that sometimes exceeded 80% (M.Ebener and J. Bence, unpublished data).

Lake whitefish size-at-age declined in Lake Michigan dur-ing the 1990s. Moreover, condition factor dropped rapidlybetween 1995 and 1998 (Fig. 10). Reduced growth rate dur-ing the 1990s may have resulted from density-dependentfactors, as sustained high levels of lake whitefish abundanceresulted in increased intraspecific competition for food re-sources. Other factors may also have been involved. For ex-ample, zebra mussel density increased substantially innorthern Lake Michigan during the same time that lakewhitefish condition declined precipitously (T. Nalepa, un-

2002 NRC Canada


Fig. 9. Annual commercial harvest of lake whitefish from LakeMichigan, 19701998.

Fig. 10. Mean condition factor, K, for age-4 (!), age-5 (), andage-6 (#) lake whitefish in northern Lake Michigan, 19851998.K= 105WL3, where Wis weight (g) and L is total length (mm).


12/18

published data). Hoyle et al. (1999) linked poor body condi-tion of lake whitefish in Lake Ontario with zebra musselinvasion of the lake; perhaps expansion of the zebra musselpopulation also had a negative effect on lake whitefish con-dition in Lake Michigan.

Burbot

Diet of adult burbot (Lota lota) in Lake Michigan hasbeen diverse and has included both fish and benthic inverte-brates, especially crayfish (Fratt et al. 1997; S. Hart and J.Jonas, Michigan Department of Natural Resources,Charlevoix, MI 49720, U.S.A., unpublished data). In deeperwaters (>39 m), deepwater sculpin has been an importantcomponent of burbot diet since the 1930s.

The burbot population in Lake Michigan was greatly re-duced in size because of sea lamprey predation during the1950s (Wells and McLain 1973). Recovery of the populationbegan, albeit very slowly, during the late 1960s and 1970s(Eck and Wells 1987). However, rapid increase in abundanceoccurred in the 1980s (Fig. 7). Thus, burbot recovery wasdelayed for more than 10 years compared with lakewhitefish recovery, even though both recoveries are attrib-uted to sea lamprey control. Eshenroder and Burnham-Curtis(1999) contended that this delay was due to alewife interfer-ence with burbot reproduction. Burbot larvae are pelagic;therefore, they are susceptible to predation by alewives. Ale-wife density may have decreased sufficiently during the1970s to allow for a strong recovery by the burbot popula-tion during the 1980s. According to bottom trawl surveysconducted by the USGSGLSC, lake-wide abundance ofburbot increased exponentially during the 1980s and thenleveled off during the 1990s (Fig. 7). As mentioned above,patterns in the long-term series for abundances, coupled withdiet data for burbot, suggested that predation by burbot hasmaintained the deepwater sculpin population at a level lowerthan that observed during the early 1980s (Figs. 6 and 7).Based on GLSC bottom trawl surveys, lake-wide biomass ofburbot in Lake Michigan proper averaged about 9.7 kt dur-ing 19901999.

Yellow perchThe yellow perch population in Lake Michigan has

yielded valuable sport and commercial fisheries (Wells andMcLain 1973). During the 1980s and 1990s, yellow perchwas the most popular sport-caught fish (Bence and Smith1999).

Decline in yellow perch abundance in Lake Michigan dur-ing the 1960s has been attributed to the alewife invasion

(Wells 1977), indicating that the levels of alewives charac-teristic of the 1960s and early 1970s were sufficiently highto interfere with yellow perch reproduction.

With strong year-classes produced in 1980 and 19831988, yellow perch abundance in Lake Michigan increasedduring the 1980s (Jude and Tesar 1985; Eck and Wells 1987;Makauskas and Clapp 2000). This recovery in the yellowperch population has been attributed to the decline in alewifeabundance during the 1970s (Jude and Tesar 1985; Eck andWells 1987). Lake-wide abundance of yellow perch re-mained relatively high during 19851990, but then de-creased during 19901995 because year-class strength wasweak during 19891993 (Makauskas and Clapp 2000). Year-

class strength remained weak during 19951997 and in1999, but was moderate in 1998. No single factor could ex-plain poor recruitment of yellow perch in Lake Michiganproper between 1989 and 1997. Lake-wide abundance ofalewife was as low during 19891997 as it was during19831988 (Fig. 4), and therefore it seems unlikely that in-terference by alewives was primarily responsible for poor

recruitment during 19891997 throughout the lake. Nonethe-less, Shroyer and McComish (2000) reported a negative re-lationship between yellow perch year-class strength andalewife abundance in Indiana waters of Lake Michigan dur-ing 19841996. Skewing of the sex ratio (more males thanfemales) during 19921996 may have partially contributedto poor recruitment (Makauskas and Clapp 2000). Skewedsex ratio in the yellow perch population may have been dueto commercial harvest operations, because females growfaster than males and the commercial fishery targets largerindividuals. Perhaps the number of female spawners was sodiminished that the probability of producing a moderate orlarge size year-class was low. During some years, slow rateof warming of nearshore waters in the spring may have par-

tially contributed to production of weak year-classes(Makauskas and Clapp 2000). Finally, data from Illinois wa-ters suggested that bottom-up effects, namely low zooplank-ton abundance during June and July, may also have beenpartly responsible for poor recruitment, presumably becausesurvival of yellow perch fry was reduced due to low foodavailability (Makauskas and Clapp 2000).

Walleye

Walleye (Stizostedion vitreum) has been one of the mostecologically and economically important fishes in Michiganwaters of Green Bay (Schneeberger 2000). Other centers ofrelatively high walleye density in Lake Michigan includedsouthern Green Bay and the Muskegon vicinity of easternLake Michigan. Walleye abundance in Lake Michigan fluc-tuated greatly during the 1900s and these fluctuations havebeen attributed to overexploitation, deterioration of waterquality (in southern Green Bay), loss of habitat (in southernGreen Bay), and interference by alewives with walleye re-production.

Walleye abundance in Green Bay declined precipitouslyduring the 1960s, but rebounded during the 1970s becauseof management actions that included stocking and imposi-tion of restrictions on commercial and recreational fishing(Schneeberger 2000). Today, the walleye population in LittleBay de Noc (northern Green Bay) supports a world-classrecreational fishery, and reduction in stocking rates is now

being considered by managers because the fishery is nearlyself-sustaining. Walleye diet in northern Green Bay has re-mained consistent between the 1960s and 1990s, with rain-bow smelt and alewives representing the bulk of the diet(Schneeberger 2000).

Lake sturgeon

Population size of lake sturgeon (Acipenser fulvescens)(for individuals 22.7 kg in weight) in Lake Michigan dur-ing 18251890 has been estimated to have been between145 000 and 2 406 000 individuals (Hay-Chmielewski andWhelan 1997). The lake sturgeon population in LakeMichigan collapsed during the early 1900s owing to over-

2002 NRC Canada



13/18

exploitation and habitat degradation (Wells and McLain1973). During the 1990s, small populations of lake sturgeonhave been associated with at least 13 tributaries (Hay-Chmielewski and Whelan 1997). With continued removal ofdams on tributaries anticipated in the near future, the size oflake sturgeon population in Lake Michigan should increaseas more suitable spawning habitat becomes available.

Emerald shinerThe emerald shiner (Notropis atherinoides) population in

Lake Michigan crashed during the early 1960s, coincidentwith buildup of the alewife population (Wells and McLain1973). Apparently, alewives fed on the pelagic eggs and lar-vae of the emerald shiner. Emerald shiner abundance has re-mained at very low levels during 19702000 (D. Jude,unpublished data; J. Dettmers, Illinois Natural History Sur-vey, Zion, IL 60099, U.S.A., unpublished data; R. Rost, Wis-consin Department of Natural Resources, Peshtigo, WI54157, U.S.A., unpublished data), suggesting that the rela-tively low alewife abundance during 19842000 was still suf-ficiently high to interfere with emerald shiner reproduction.

Round gobyThe round goby, an invader from the Ponto-Caspian re-

gion, established populations in Calumet Harbor (southernLake Michigan) and in several harbors on both the east andwest sides of Lake Michigan including Milwaukee, SturgeonBay, Escanaba, Grand Haven, Muskegon, and Charlevoixduring the 1990s (Clapp et al. 2001). Zebra mussels form asubstantial portion of the diet of round gobies in the GreatLakes (C. Knight, Ohio Department of Natural Resources,Fairport, OH 44077, U.S.A., personal communication; M.Thomas, Michigan Department of Natural Resources, Mt.Clemens, MI 48045, U.S.A., personal communication). In

turn, round gobies are preyed upon by several species offish. The precipitous decline of mottled sculpin (Cottusbairdi) abundance in Calumet Harbor, as well as decreasesin abundances of johnny darter (Etheostoma nigrum) andlogperch (Percina caprodes) in Lake St. Clair, has beenlinked to the round goby invasion (Janssen and Jude 2001;M. Thomas, Michigan Department of Natural Resources,Mt. Clemens, MI 48045, U.S.A., personal communication).It remains to be seen whether these seemingly negative ef-fects of round gobies on sculpins and darters will be con-fined to shallow and protected waters of Lake Michigan.

Lake herringLake herring (Coregonus artedi) was once a prominent

member of the fish community in Lake Michigan, and thelake herring population once supported a valuable commer-cial fishery (Wells and McLain 1973). The crash in the lakeherring population during the 1950s has been attributed tooverexploitation and interference by rainbow smelt with lakeherring reproduction (Wells and McLain 1973). Althoughthere have been no signs of a lake-wide recovery, a recentincrease in lake herring abundance in Grand Traverse Bay ofnorthern Lake Michigan has been observed. Sport harvest oflake herring in Grand Traverse Bay indicated the presence oflake herring at very low levels in 1985 and 1993, but sportharvest increased three- to seven-fold during 19951998compared with the 1993 harvest (G. Rakoczy, Michigan De-

partment of Natural Resources, Charlevoix, MI 49720,U.S.A., personal communication).

Contaminants and global climate change

stressors

ContaminantsThe issue of contaminants acting as stressors to GreatLakes biota emerged in the 1970s (Evans 1988). Conse-quently, the governments of the United States and Canadanegotiated the Great Lakes Water Quality Agreement in1972, and revised this agreement in 1978. This agreementled to many remedial actions to improve the quality of GreatLakes waters. Additionally, production of polychlorinatedbiphenyls (PCBs) was banned, use of the pesticide DDT(1,1,1-trichloro-2,2-bis[p-chlorophenyl]ethane) was discon-tinued, and uses of the organochlorine pesticides chlordaneand dieldrin were restricted in both the United States andCanada during the 1970s.

As a consequence of the abovementioned measures, con-

centrations of most contaminants in Lake Michigan biota de-creased substantially between 1970 and 2000. For example,total PCB concentration in Lake Michigan zooplankton de-creased between 1982 and 1995 (A. Trowbridge and D.Swackhamer, University of Minnesota, Minneapolis, MN55455, U.S.A., personal communication). Total PCB con-centration in Mysis, Diporeia, prey fishes, and salmonines ofLake Michigan decreased by roughly an order of magnitudebetween 1975 and 1995 (DeVault et al. 1996; Madenjian etal. 1998b, 1998c). During the 1970s and early 1980s, de-creases were observed in total DDT, dieldrin, andoxychlordane concentrations in lake trout (DeVault et al.1996), and in PCB, oxychlordane, DDT, and dieldrin con-centrations in herring gull (Larus argentatus) eggs (Pekarik

and Weseloh 1998). In contrast, toxaphene concentrations inlake trout did not decrease between 1986 and 1992 (DeVaultet al. 1996).

Very few lake-wide temporal changes in Lake Michiganbiota between 1970 and 2000 have been shown to be directlylinked to stressor effects exerted by contaminants. The dra-matic increase in double-crested cormorant (Phalacrocoraxauritus) population size during the late 1980s and 1990s wasdue, at least in part, to the substantial reduction of contami-nant levels within the lake ecosystem during the 1970s(Weseloh et al. 1995). Number of double-crested cormorantnests on islands within Lake Michigan (including GreenBay) increased exponentially from 75 nests in 1977 to 4743nests during 19891990, and then to 28 158 nests in 1997

(F.J. Cuthbert, University of Minnesota, St. Paul, MN 55108,U.S.A., personal communication). Weseloh et al. (1983)documented shell thinning with high egg breakage and highegg loss (95%) with concomitant reproductive failure inGreat Lakes cormorants during the early 1970s, and also re-ported high DDE (a metabolite of DDT) concentrations incormorant eggs. Apparently, high concentration of DDE inthe cormorants was causing eggshell thinning and reproduc-tive failure. Increased predation by the exploding populationof double-crested cormorants in Lake Michigan may havecaused a reduction in abundance of smallmouth bass(Micropterus dolomieui) in the vicinity of Beaver Island innorthern Lake Michigan; however, research results to date

2002 NRC Canada



14/18

are not conclusive (D. Peterson, Central Michigan Univer-sity, Mt. Pleasant, MI 48859, U.S.A., personal communica-tion). To the best of our knowledge, no lake-wide temporalchanges in phytoplankton communities, benthic invertebratecommunities, zooplankton communities, or fish communi-ties in Lake Michigan have been shown to be directly attrib-utable to contaminant effects (Evans 1988). Nevertheless,

contaminants may have influenced health of brown bullhead(Ameiurus nebulosus) populations in certain tributaries(Baumann et al. 1996).

Global climate change

An increase in atmospheric carbon dioxide enhances heatretention (Regier et al. 1990). Burning fossil fuels andchanges in land use are correlated with an increase in carbondioxide concentration in the earths atmosphere, and there-fore global climate change likely represents ananthropogenic stressor to the global environment. Althoughnot all climatologists agree as to the nature, extent, intensity,and timing of the effects of increased carbon dioxide contentin the earths atmosphere, there is general consensus that cli-mate warming will occur (Regier et al. 1990; Doak and Mor -ris 1999).

Magnuson et al. (1997) predicted the effects of changes inthermal regimes associated with global climate change onprimary production, zooplankton biomass, maximum sus-tainable fishery yield, growth and consumption by fish, fishreproduction and overwinter survival, success of invasionsby exotic fish species, and dissolved oxygen concentration inthe Great Lakes. For Lake Michigan, McCormick (1990)used computer simulation modeling to explore potentialchanges in thermal structure and cycle of Lake Michigan un-der three different scenarios for global climate change.Using predictions from McCormick (1990), Magnuson et al.

(1990) forecasted that thermal habitat within Lake Michiganwill increase under global warming scenarios for fish speciesfrom all thermal guilds. Additionally, Magnuson et al.(1997) reported that global warming has been predicted toeventually cause a 0.99- to 2.48-m decrease in Lake Michi-gans water level; however, these researchers did not specifi-cally address effects of this lowering of lake level on LakeMichigan biota.

Based on available data, we are unable to ascribe any ofthe lake-wide changes in the Lake Michigan food web dur-ing 19702000 to global climate change, mainly because in-sufficient data are available to establish long-term, lake-widetrends in water temperature characteristics in Lake Michiganor to clearly show that a long-term trend in water tempera-

ture characteristics is linked to global climate change. Long-term trends in water temperatures were not consistent be-tween two sites in Lake Michigan, including one site at St.Joseph, along the southeastern shore of Lake Michigan, andanother at the southern end of Green Bay (McCormick andFahnenstiel 1999). Moreover, even though the duration ofsummer thermal stratification increased significantly at St.Joseph during 19601992, McCormick and Fahnenstiel(1999) cautioned that the series did not extend far enoughback in time to resolve whether the trend was part of a natu-ral cycle or was caused by anthropogenic activity. Finally,recent low water levels in Lake Michigan can be explainedby quasi-periodicity in water level fluctuations (periods of

roughly 30 and 150 years) and by unusually low rainfall dur-ing 1998 and 1999, without evoking a global climate changeeffect (T. Thompson, Indiana Geological Survey,Bloomington, IN 47405, U.S.A., personal communication).

Discussion

Sea lamprey control and top-down effectsWhereas Wells and McLain (1973) concluded that sea

lamprey and alewife invasions had a greater effect on theLake Michigan fish community than any of the otheranthropogenic stressors during 19401970, control of sealamprey and alewife populations was paramount in effectingchanges within the Lake Michigan ecosystem during 19702000. Spectacular recovery of the lake whitefish populationin Lake Michigan between 1965 and 2000 was partially at-tributable to sea lamprey control. Further, control of sea lam-prey, beginning in the 1950s, enabled the successful stockingeffort launched during the 1960s to increase salmonine bio-mass. In turn, salmonine populations, especially the chinook

salmon population, were primarily responsible for the de-cline in alewife abundance during the 1970s and early1980s. Strong recovery of burbot during the 1980s was prob-ably fueled by control of both sea lamprey and alewife.Moreover, recovery of deepwater sculpin during the 1970swas most likely facilitated by reduction in alewife density.Also, relaxation of the interference effect by alewives onyellow perch reproduction is the most plausible explanationfor the rapid increase in yellow perch abundance during the1980s. Finally, the sudden increase in abundance of the verylarge-bodied cladoceran D. pulicaria during the 1980s wasprobably also due to reduction in alewife abundance.

Evidence that predation by salmonines drove the reduc-tion in alewife abundance is strong. Decline in alewife abun-

dance during the 1970s and early 1980s was wellsynchronized with buildup of salmonine populations. More-over, the period of relatively low alewife abundance between1984 and 2000 coincided with high predation by salmonineson alewives, as estimated by our modeling. Alewives havedominated salmonine diets in Lake Michigan from the 1960sthrough the 1990s. Additionally, annual alewife abundancesince the early 1970s was much better correlated with esti-mates of annual consumption of alewives by salmoninesthan with cold temperatures. Based on our estimates of ale-wife consumption by salmonines, the commercial fishery foralewives was a relatively minor contributor to the observeddecrease in alewife abundance. Annual commercial harvestof alewives from Lake Michigan averaged about 15 kt dur-

ing 19731984 (S. Nelson, USGSGLSC COMCAT data-base, Ann Arbor, MI 48105, U.S.A.), whereas annualconsumption of alewives by salmonines averaged about 60kt over the same time period. The Lake Michigan commer-cial fishery for alewives was closed in 1991 and has notbeen reopened.

Intensive stocking of salmon and trout into Lake Michigancan be viewed as a manipulation of the ecosystem. Lake-wide monitoring of prey fish populations has provided con-siderable insight into predatorprey dynamics within theLake Michigan ecosystem. Predatorprey interaction be-tween chinook salmon and alewife has been one of thestrongest trophic links within the ecosystem since the 1970s.

2002 NRC Canada



15/18

Continued monitoring of prey fish populations, as well assurveillance of other components of the ecosystem, shouldprovide further understanding of the Lake Michigan foodweb. Such long-term studies are especially appropriate forecosystem manipulations (Pace and Cole 1989).

Bottom-up effects

Several changes within the Lake Michigan ecosystem dur-ing 19702000 may be attributable to bottom-up effects. De-cline in abundance of all three dominant benthicmacroinvertebrate groups, including Diporeia, oligochaetes,and sphaeriids, between 1980 and 1993 in nearshore (50 mdeep) waters of Lake Michigan was believed to be chieflydue to a decrease in primary production in nearshore waters.Although long-term data for spring total phosphorus, an in-dicator of primary production, in nearshore waters of LakeMichigan were not available, comparison of measurementsduring 19982000 with measurements from the early 1970ssuggested a decrease in primary production over this timeperiod. In contrast, long-term data were readily available foroffshore waters, and these data showed no long-term trend inprimary production during 19731998. Similarly, abun-dances of the three major groups of benthicmacroinvertebrates in offshore waters did not trend upwardor downward during this time period. The apparent declinein primary production in nearshore waters of Lake Michiganbetween the 1970s and 19982000 was likely due to the de-crease in phosphorus loadings between 1980 and 1987. Thecontinued decrease in Diporeiaabundance during the 1990sin nearshore waters of southern and southeastern Lake Mich-igan coincided with the zebra mussel invasion, although aspecific mechanism whereby zebra mussels may have re-duced Diporeia abundance remains unidentified.

Have bottom-up effects, including the apparent reduction

in nearshore primary production and the continued declineof Diporeia, already precipitated a decrease in abundancesof important fish populations in Lake Michigan? No clear ef-fects are evident yet. The decrease in bloater abundance dur-ing the 1990s was attributed to an extended period of verylow recruitment, which may be related to intraspecific mech-anisms rather than bottom-up effects. Lake-wide abundancesof alewife, salmonines, burbot, lake whitefish, and sculpinshave not decreased during the 1990s. Even though data fromIllinois waters suggested that bottom-up effects may havepartly contributed to a prolonged period of low yellow perchrecruitment, available data were insufficient to support thecontention that a long-term decrease in zooplankton abun-dance was primarily responsible for the extended period of

low recruitment. Rainbow smelt abundance decreased during19921999; however, explanations for trends in rainbowsmelt abundance, as determined by GLSC bottom trawl sur-veys over the past 30 years, are not obvious. Eck and Wells(1987) also experienced difficulty in interpreting long-termtrends in rainbow smelt abundance.

Have these bottom-up effects already influenced growthrates or condition of Lake Michigan fishes? The decrease inlake whitefish condition between 1995 and 1998 may havebeen partially due to some bottom-up effects. An increase inproportion of zebra mussels or a decrease in proportion ofDiporeia in lake whitefish diet could have led to reducedlake whitefish condition. The recent decline in alewife con-

dition may be due to bottom-up effects. Bloater growth ratedeclined between 1973 and 1994, during which time thepopulation increased tremendously in size. However, bloatergrowth rate has actually increased during 19941999, as thepopulation has undergone a drastic decrease in size. To de-tect future impacts of bottom-up effects on fish populations,surveillance of fish abundance, growth rates, lipid content,

and condition should be continued. Zebra mussel abundancein Lake Michigan has not yet peaked, and the quagga mussel(Dreissena bugensis) invasion of Lake Michigan has just be-gun (Nalepa et al. 2001).

Overexploitation of fish stocksOverall, overexploitation of fish populations appeared to

play a relatively minor role in shaping the Lake Michiganfood web during 19702000. Commercial harvest of yellowperch during the 1990s may have contributed to a skewingof the sex ratio (relatively low number of female spawners),which may have contributed to prolonging the period of lowrecruitment to the yellow perch population. Yellow perch re-cruitment remained low during 19891997. Consequently,

the yellow perch commercial fishery in Lake Michigan wasclosed by 1998.

Bythotrephes and round goby invasionsAlthough invasions by Bythotrephesand round goby have

caused some changes to the Lake Michigan food web duringthe late 1980s and 1990s, the earlier invasions by sea lam-prey and alewife had more serious consequences for theLake Michigan fish community. Through predation,Bythotrephes may have reduced the abundance of certainsmaller cladocerans. However, the impact of this reductionon the rest of the Lake Michigan ecosystem remains unde-fined. Total phytoplankton biomass in offshore waters has

exhibited no long-term trends since the early 1970s. Al-though species composition of the zooplankton communityhas undergone substantial changes, total zooplankton bio-mass in offshore waters of Lake Michigan has shown nolong-term trends since the 1970s. Abundance of Mysis inoffshore waters has not changed appreciably since the1970s. Alewife abundance has neither increased nor de-creased since the time of the Bythotrephes invasion, andsalmonine biomass has remained relatively high throughoutthe occupation of the lake by Bythotrephes. Moreover,Bythotrephes has been a relatively minor component of thediets of the prey fish and salmonine communities in LakeMichigan since the late 1980s (Davis et al. 1997; R. Elliott,U.S. Fish and Wildlife Service, Green Bay, WI 54311,

U.S.A., personal communication). The round goby invasionwas linked to the decline of mottled sculpin in Calumet Har-bor. To date, however, no lake-wide changes in abundance ofany Lake Michigan biota have been ascribed to the roundgoby invasion.

Contaminants and global climate change stressorsAlthough bans on use of pesticides and PCBs have led to

dramatic declines in contaminant concentrations of LakeMichigan biota between 1970 and 2000, very few lake-widechanges in abundances of Lake Michigan biota have beenshown to be directly attributable to stressor effects from con-taminants. Rapid increase in the double-crested cormorant

2002 NRC Canada



16/18

2002 NRC Canada


population associated with Lake Michigan during the late1980s and 1990s was partially due to substantial reductionof contaminant levels within the lake ecosystem during the1970s (Weseloh et al. 1995). To date, lake-wide conse-quences of this increase in cormorant abundance on the restof the lake ecosystem remain unclear.

We were unable to attribute any of the lake-wide changes

in the Lake Michigan food web between 1970 and 2000 toglobal climate change effects. That is not to say that globalclimate change has not already had some effect on the LakeMichigan ecosystem. Rather, time series for water tempera-ture characteristics are not sufficiently long to resolvewhether recent trends were part of a natural cycle or wereforced by anthropogenic activity. Furthermore, insufficientdata were available to identify a common long-term trend inwater temperature characteristics across several locations inthe lake. We would expect that ecological changes associ-ated with global climate change would occur much moreslowly than changes induced by invasions of exotic species.

Other effects

Finally, we would like to point out that one of the majorchanges occurring in the Lake Michigan food web between1970 and 2000 may not have been driven by any of theanthropogenic stressors mentioned above. Rise in bloaterabundance during the 1980s and fall in bloater abundanceduring the 1990s may simply have represented natural oscil-lations in bloater population size. In other words, bloaterabundance may exhibit quasi-regular natural cycles, with aperiod of approximately 30 years. Through density-dependent mechanisms, bloaters may be regulating their ownabundance, largely independent of human interventions orinteractions with other fishes. Given that bloater longevity inLake Michigan is at least 12 years (G. Fleischer, unpub-lished data), a 30-year cycle in their abundance would be in

accord with theory developed by Nisbet (1997) for delayed-feeback population cycles.

Acknowledgments

This paper would not have been possible without the ef-forts of the Lake Michigan Technical Committee, includingchairperson M. Holey, as well as D. Clapp, B. Eggold, R.Hess, J. Kubisiak, S. Marcquenski, E. Olson, M. Toneys, andG. Wright. We thank L. Wells, senior author of the SCOL-1paper for Lake Michigan, for providing helpful comments.M. Hansen and S. Kerr also reviewed the manuscript. Withthis paper, we pay tribute to the memory of Heang T. Tin, apioneer of larval fish ecology in Lake Michigan. This work

was supported in part by Michigan Sea Grant College Pro-gram grants R/FM-1 and R/GL-46 to J.R.B. This article isContribution 1172 of the USGSGLSC, GLERL contribu-tion No. 1224, and CGLAS contribution No. 622. This pub-lication is a result of work sponsored by the CooperativeInstitute for Limnology and Ecosystems Research undercooperative agreement NA67RJ0148 from the Office of Oce-anic and Atmospheric Research, National Oceanic and Atmo-spheric Administration, U.S. Department of Commerce.

References

Baumann, P.C., Smith, I.R., and Metcalfe, C.D. 1996. Linkages be-

tween chemical contaminants and tumors in benthic Great Lakesfish. J. Gt. Lakes Res. 22: 131152.

Bence, J.R., and Smith, K.D. 1999. An overview of recreationalfisheries of the Great Lakes. In Great Lakes fisheries and policymanagement: a binational perspective. Edited by W.W. Taylorand C.P. Ferreri. Michigan State University Press, East Lansing,Mich. pp. 259306.

Benjamin, D.M. 1998. Chinook salmon (Oncorhynchustshawytscha) population dynamics in Lake Michigan, 19851996. M.Sc. thesis, Michigan State University, East Lansing.

Benjamin, D.M., and Bence, J.R. 2002. Statistical catch-at-ageanalysis of chinook salmon in Lake Michigan, 19851996. Fish-eries Research Report, Fisheries Division, Michigan Departmentof Natural Resources, Ann Arbor.

Bennet, E.B. 1986. The nitrifying of Lake Superior. Ambio, 15:272275.

Brooks, A.S., and Torke, B.G. 1977. Vertical and seasonal distribu-tion of chlorophylla in Lake Michigan. J. Fish. Res. Board Can.34: 22802287.

Brown, E.H., Jr. 1972. Population biology of alewives, Alosapseudoharengus, in Lake Michigan, 194970. J. Fish. Res.Board Can. 29: 477500.

Chang, W.Y.B., and Rossman, R. 1988. Changes in the abundanceof blue-green algae related to nutrient loadings in the nearshorezone of Lake Michigan. Hydrobiologia, 157: 271278.

Clapp, D.F., Schneeberger, P.J., Jude, D.J., Madison, G., and Pistis,C. 2001. Monitoring round goby (Neogobius melanostomus)population expansion in eastern and northern Lake Michigan. J.Gt. Lakes Res. 27: 335341.

Davis, B.M., Savino, J.F., and Ogilvie, L.M. 1997. Diets of foragefish in Lake Michigan. U.S. Environmental Protection AgencyReport EPA/IAG DW14947692-01-0, Chicago, Ill.

DeVault, D.S., Hesselberg, R., Rodgers, P.W., and Feist, T.J. 1996.Contaminant trends in lake trout and walleye from the Lauren-tian Great Lakes. J. Gt. Lakes Res. 22: 884895.

Doak, D.F., and Morris, W. 1999. Detecting population-level con-sequences of ongoing environmental change without long-termmonitoring. Ecology, 80: 15371551.

Eck, G.W., and Brown, E.H., Jr. 1985. Lake Michigans capacity tosupport lake trout (Salvelinus namaycush) and other salmonines:an estimate based on the status of prey populations in the 1970s.Can. J. Fish. Aquat. Sci. 42: 449454.

Eck, G.W., and Wells, L. 1987. Recent changes in Lake Michigansfish community and their probable causes, with emphasis on therole of the alewife (Alosa pseudoharengus). Can. J. Fish. Aquat.Sci. 44(Suppl. 2): 5360.

Eshenroder, R.L., and Burnham-Curtis, M.K. 1999. Species suc-cession and sustainability of the Great Lakes fish community. InGreat Lakes fisheries and policy management: a binational per-spective. Edited by W.W. Taylor and C.P. Ferreri. Michigan

State University Press, East Lansing. pp. 145184.Evans, M.S. 1988. Toxic contaminants and ecosystem health: a

Great Lakes focus. John Wiley & Sons, New York.Evans, M.S. 1992. Historic changes in Lake Michigan zooplankton

community structure: the 1960s revisited with implications fortop-down control. Can. J. Fish. Aquat. Sci. 49: 17341749.

Evans, M.S., and Jude, D.J. 1986. Recent shifts in Daphniacommu-nity structure in southeastern Lake Michigan: a comparison of theinshore and offshore regions. Limnol. Oceanogr. 31: 5667.

Fahnenstiel, G.L., and Scavia, D. 1987a. Dynamics of Lake Michi-gan phytoplankton: recent changes in surface and deep commu-nities. Can. J. Fish. Aquat. Sci. 44: 509514.

Fahnenstiel, G.L., and Scavia, D. 1987b. Dynamics of Lake Michi-


17/18

gan phytoplankton: primary production and growth. Can. J.Fish. Aquat. Sci. 44: 499508.

Fahnenstiel, G.L., and Scavia, D. 1987c. Dynamics of Lake Michi-gan phytoplankton: the deep chlorophyll layer. J. Gt. Lakes Res.13: 285295.

Fahnenstiel, G.L., Krause, A.E., McCormick, M.J., Carrick, H.J., andSchelske, C.L. 1998. The structure of the planktonic food-web inthe St. Lawrence Great Lakes. J. Gt. Lakes Res. 24: 531554.

Fitzsimons, J.D., Brown, S.B., Honeyfield, D.C., and Hnath, J.G.1999. A review of early mortality syndrome (EMS) in GreatLakes salmonids: relationship with thiamine deficiency. Ambio,28: 915.

Fratt, T.W., Coble, D.W., Copes, F., and Brusewitz, R.E. 1997. Dietof burbot in Green Bay and western Lake Michigan with com-parison to other waters. J. Gt. Lakes Res. 23: 110.

Gardner, W.S., Nalepa, T.F., Frez, W.A., Cichocki, E.A., andLandrum, P.F. 1985. Seasonal patterns in lipid content of LakeMichigan macroinvertebrates. Can. J. Fish. Aquat. Sci. 42:18271832.

Hatch, R.W., Haack, P.M., and Brown, E.H., Jr. 1981. Estimationof alewife biomass in Lake Michigan, 19671978. Trans. Am.Fish. Soc. 110: 575584

Hay-Chmielewski, E.M., and Whelan, G.E. 1997. Lake sturgeonrehabilitation strategy.Special Report 18, Fisheries Division,Michigan Department of Natural Resources, Ann Arbor.

Holey, M.E., Rybicki, R.W., Eck, G.W., Brown, E.H., Jr., Marsden,J.E., Lavis, D.S., Toneys, M.L., Trudeau, T.N., and Horrall,R.M. 1995. Progress toward lake trout restoration in Lake Mich-igan. J. Gt. Lakes Res. 21(Suppl. 1): 128151.

Hoyle, J.A., Schaner, T., Casselman, J., and Dermott, R. 1999.Changes in lake whitefish (Coregonus clupeaformis) stocks ineastern Lake Ontario following Dreissena mussel invasion. Gt.Lakes Res. Rev. 4(2): 510.

Janssen, J., and Jude, D.J. 2001. Recruitment failure of mottledsculpinCottus bairdiin southern Lake Michigan, induced by thenewly introduced round goby Neogobius melanostomus. J. Gt.Lakes Res. 27: 319328.

Johengen, T.H., Johansson, O.E., Pernie, G.L., and Millard, E.S.1994. Temporal and seasonal trends in nutrient dynamics andbiomass measures in Lakes Michigan and Ontario in response tophosphorus control. Can. J. Fish. Aquat. Sci. 51: 25702578.

Jones, M.L., Koonce, J.F., and OGorman, R. 1993. Sustainabilityof hatchery-dependent salmonine f

dynamics of the lake michigan food web1970-2000

Documents